Specs

Core module

Computing unit

  • CPU

  • Dual-core 64-bit RISC-V CPU, base frequency: 400MHz, each with independent single-precision FPU

  • NPU

  • Neural network accelerator, supports operator acceleration such as convolution, pooling, and activation, computing power: 0.8Tops

  • SRAM

  • 8MB 64-bit

  • Flash

  • 16MB NOR Flash

AI performance

  • Task type

  • Supports machine vision - target detection tasks

  • Toolchain

  • All-in-one visual AI training platform, allows beginners to easily train their own models

  • Neural network model

  • Customized and improved YOLOv2

  • Model input

  • 320x240RGB, 16-bit

  • Model calculation time

  • <9ms

Camera

  • Sensors

  • 1/4-inch, color, global exposure, 60fps

  • Output image

  • Resolution: 320x240, 16-bit color depth

  • Lens

  • FOV: 40°, fixed focal length: 8mm

  • Lens holder

  • Manually adjustable pitch axis

General

  • Weight

  • 9g

  • Dimensions

  • 54x32x11mm

  • Power supply

  • Micro USB power supply or 14pin expansion port 5V power supply, supports hot swapping, short circuit protection, overcurrent protection, and overheating protection

  • Power consumption

  • About 1.25W

  • Output port

  • Serial port (inference results), SPI (video stream)

  • Auxiliary functions

  • Button x1, RGB LED x1

Touchscreen

Touchscreen

  • Screen size

  • 2.4 inches

  • Display image

  • Resolution: 320x240, 16-bit color

  • Display port

  • 8-bit MCU parallel interface

General

  • Weight

  • 42g

  • Dimensions

  • 67x49x15mm

  • Power supply

  • 5V, supports hot swapping

  • Power consumption

  • About 250mW

  • Auxiliary functions

  • Button x2, SD card slot x1 (must be used with core module)

Universal expansion module

General

  • Dimensions

  • 58x46x15mm

  • Weight

  • 22.5g

  • Power supply

  • 5V/12V, supports hot swapping

  • Power consumption

  • About 500mW

  • MCU

  • i.MX RT1064, base frequency: 600MHz, 1MB RAM + 4MB ROM

  • Open source

  • SDK development supported

  • Expandability

  • 14-pin extended interface (I2C, UART, SPI, GPIO, PWM, and power supply)

  • 100-pin expanded interface (power supply, GPIO, and other reusable functions)

  • Auxiliary functions

  • Button x2, SD slot x1

AI training platform

Software

  • No. of labeling tasks

  • Advanced Edition:3
    Professional Edition:5
    Professional Edition - AI Service Package:15

  • No. of datasets

  • Advanced Edition:3
    Professional Edition:8
    Professional Edition - AI Service Package:25

  • No. of datasets per training round

  • Advanced Edition:3
    Professional Edition:5
    Professional Edition - AI Service Package:8

  • No. of algorithms for training

  • Advanced Edition:3
    Professional Edition:5
    Professional Edition - AI Service Package:15

  • No. of simultaneous algorithm training tasks

  • Advanced Edition:1
    Professional Edition:2
    Professional Edition - AI Service Package:3

  • No. of models

  • Advanced Edition:3
    Professional Edition:8
    Professional Edition - AI Service Package:25

  • No. of algorithm applications

  • Advanced Edition:3
    Professional Edition:5
    Professional Edition - AI Service Package:15

  • Avg. algorithm training time

  • Advanced Edition:Long
    Professional Edition:Long
    Professional Edition - AI Service Package:Shorter